AMD Phenom II Review Thread

[Miss Banana]

You haven't' presented the slightest evidence that his numbers are off or that the benchmarks he used were affected by the memory difference.The burden of proof is on you.

FYI , reviews aren't made exclusively for XtremeSystems users ; I'm sure most of the people who visit review sites don't know what RAM timings are.People will buy a prebuilt system that probably has the cheapest mobo and cheapest RAM , made on a ship between ASIA and Europe/US.They saw the CPU sticker or asked for a system built around a specific CPU.

In the real world nobody cares about DDR3 2000 CL 6 , 1.8V Vcore , QPI voltage , etc.

I really do not expect you, with your ridiculous fanboyish sigs, to understand why objectivity is an important part of reviewing hardware, when it's about AMD.

To respond to your points, what kind of evidence can I possibly give, other than that Kyle's results differ from the results of websites that did bother to test in a fair way, that will make you, of all people, satisfied?

To me it's quite ridiculous that after complaining about a review that is obviously screwed up in a few serious ways, starting with unfair hardware setup, you now believe it is my burden to give proof that this unfair testing platform influenced the results.

Here you go:
http://ixbtlabs.com/articles3/cpu/am...r2-533-p2.html

If you need further "proof", you can use google.

[keithlm]

You haven't' presented the slightest evidence that his numbers are off or that the benchmarks he used were affected by the memory difference.The burden of proof is on you.

Actually there is no need to prove anything when it is something obvious that anybody with common sense can ascertain.

I have noticed that it is often a common forum tactic to tell someone that they have the "burden of proof" when it is something that is obvious and there is absolutely nothing available to counter their argument.

EDIT: Gee... why don't we use DDR3-1066 ram for all i7 benchmark reviews. You think that might be acceptable? At 9-9-9 timings.

[roofsniper]

In which cases does DDR800 -> DDR1066 have great impacts on performance? And what are those great impacts? 1-2% difference?

Oh wait, here in XS, if CPU_A is 3 % faster than CPU_B, then CPU_B is being "completely destroyed" by CPU_A. Wtf people, get the cheaper one and be happy.

yea but also since this is xs people want the most performance possible. most don't even care about the cost. so even tho i wouldn't say cpu A destroys cpu B i could understand why people want it. too bad for the normal consumer that price/perf means more than 3% performance.

[savantu]

I really do not expect you, with your ridiculous fanboyish sigs, to understand why objectivity is an important part of reviewing hardware, when it's about AMD.

To respond to your points, what kind of evidence can I possibly give, other than that Kyle's results differ from the results of websites that did bother to test in a fair way, that will make you, of all people, satisfied?

To me it's quite ridiculous that after complaining about a review that is obviously screwed up in a few serious ways, starting with unfair hardware setup, you now believe it is my burden to give proof that this unfair testing platform influenced the results.

Here you go:
http://ixbtlabs.com/articles3/cpu/am...r2-533-p2.html

If you need further "proof", you can use google.

This has to be the most strange reply ever ; you reply by giving a link which supports my argument ? I don't know if to LMAO or shrug.


The difference between DDR2 533 and DDR2 1066 on Phenom 2 is 3%.

What's the difference between DDR2 800 as used by Kyle and DDR 2 1066 ? 0.5% ? Is it above variability ?

I'll end with the conclusion from the review

There is even no need to comment on the last fact. I remind you that memory frequencies differed twofold. Twofold difference on one hand, and 3% performance gain on the other hand - do you need any comments?

Actually there is no need to prove anything when it is something obvious that anybody with common sense can ascertain.

I have noticed that it is often a common forum tactic to tell someone that they have the "burden of proof" when it is something that is obvious and there is absolutely nothing available to counter their argument.

EDIT: Gee... why don't we use DDR3-1066 ram for all i7 benchmark reviews. You think that might be acceptable? At 9-9-9 timings.

Looks like your "common sense" isn't as common sense since I have a review generously linked by Banana man that points out how little memory speed matters with Phenom 2.

DDR2 533 to DDR 2 1066 and you gain 3% overall ? Common sense ?!

[keithlm]

Looks like your "common sense" isn't as common sense since I have a review generously linked by Banana man that points out how little memory speed matters with Phenom 2.

DDR2 533 to DDR 2 1066 and you gain 3% overall ? Common sense ?!

Ah... but that is a 3% average.

Considering there are some percentages in the double digits... one would need to be blind or biased to completely dismiss the difference and blatantly use the slower ram in a professional benchmark review.

(Oh and you claim 3% for 533 to 1066 and then you claim 0.5% for 800 to 1066. Do you have PROOF of that? I think NOT.)

[Calmatory]

yea but also since this is xs people want the most performance possible. most don't even care about the cost. so even tho i wouldn't say cpu A destroys cpu B i could understand why people want it. too bad for the normal consumer that price/perf means more than 3% performance.

If that was the case, everyone would be having i7 builds and AMD would be unknown for most of the people.

[roofsniper]

Ah... but that is a 3% average.

Considering there are some percentages in the double digits... one would need to be blind or biased to completely dismiss the difference and blatantly use the slower ram in a professional benchmark review.

(Oh and you claim 3% for 533 to 1066 and then you claim 0.5% for 800 to 1066. Do you have PROOF of that? I think NOT.)

if it would scale linearly it would be a 1.5% advantage.

[Miss Banana]

This has to be the most strange reply ever ; you reply by giving a link which supports my argument ? I don't know if to LMAO or shrug.


The difference between DDR2 533 and DDR2 1066 on Phenom 2 is 3%.

Very good! You can talk big.
Now the next step is to try to actually read the article I linked...
Pay special attention the gaming benchmarks.

Also try not to miss the conclusion this time.
The biggest performance gain in applications: Unreal Tournament 3 (22%), World in Conflict (10%), 7-Zip (10%).

Not bad eh?
And this is about the OLD Phenom. Since the new phenom is not only faster clockspeed wise, but also ipc wise, it only makes sense the processor would need more bandwidth this time around yes?

No actually I think I am explaining something to someone who does not really want to understand to begin with.

[Shintai]

Funny how the posts was when hardocp reviewed conroe. Anyway, without talking about hardocp and its junk as such. Those memory+speed wouldnt change anything at all. Less than 1%. 0.5% is a good bet on average.

Very good! You can talk big.
Now the next step is to try to actually read the article I linked...
Pay special attention the gaming benchmarks.

Also try not to miss the conclusion this time.
[b] The biggest performance gain in applications: Unreal Tournament 3 (22%), World in Conflict (10%), 7-Zip (10%).

Not bad eh?
And this is about the OLD Phenom. Since the new phenom is not only faster clockspeed wise, but also ipc wise, it only makes sense the processor would need more bandwidth this time around yes?

No actually I think I am explaining something to someone who does not really want to understand to begin with.

Actually faster memory is better for the "old" phenom. I think cache blinds you in terms of "IPC" gains. On a pure core view I dont think Phenom 2 is any faster than Phenom 1.

Its abit like a Core 2 with its large cache. DDR2-667 vs DDR3-2000 basicly makes no difference. Tho you could claim it FSB limited. But we already seen the 1066vs1333FSB....

Anyway, look on the scores. I would say UT2 and maybe World in conflict is simply bad testing.

Call of duty..0%. CoH 2%, CoJ 2%. Stalker 4%, Crysis 5%.

Also a fun note. DDR2-800 in the end might not have been so bad...

http://www.anandtech.com/printarticle.aspx?i=2989

Phenom 2 used Corsair DDR2-8888. Thats 4-4-4-12 at DDR2-1100. I´m sure you can have some pretty good timings at DDR2-800 speed as in 3-3-3? The DDR3 is 8-8-8 I think

[justapost]

So does "doesn't suck" count?

http://uk.youtube.com/watch?v=hM4oCe...eature=related

[Miss Banana]

Funny how the posts was when hardocp reviewed conroe. Anyway, without talking about hardocp and its junk as such. Those memory+speed wouldnt change anything at all. Less than 1%. 0.5% is a good bet on average.



Actually faster memory is better for the "old" phenom. I think cache blinds you in terms of "IPC" gains. On a pure core view I dont think Phenom 2 is any faster than Phenom 1.

Its abit like a Core 2 with its large cache. DDR2-667 vs DDR3-2000 basicly makes no difference. Tho you could claim it FSB limited. But we already seen the 1066vs1333FSB....

Anyway, look on the scores. I would say UT2 and maybe World in conflict is simply bad testing.

Call of duty..0%. CoH 2%, CoJ 2%. Stalker 4%, Crysis 5%.

Also a fun note. DDR2-800 in the end might not have been so bad...

http://www.anandtech.com/printarticle.aspx?i=2989

Phenom 2 used Corsair DDR2-8888. Thats 4-4-4-12 at DDR2-1100. I´m sure you can have some pretty good timings at DDR2-800 speed as in 3-3-3? The DDR3 is 8-8-8 I think

Do you have any proof or reason to think this whatsoever?

[Shintai]

Do you have any proof or reason to think this whatsoever?

You can start read the link I provided. Plus your own link actually contradicts you.

[Miss Banana]

You can start read the link I provided. Plus your own link actually contradicts you.

Your link has little to do with the subject, and my own link shows a good improvement in some benchmarks as memory speed increases.

[Shintai]

Anyway I do love the lost planet run. The demonstration between the past and the future. Games like supreme commander would look the same. Plus the overall graphs in all resolutions.

Core 2 and Phenom 2 is the past. AMDs graphics department might actually be its CPU departments biggest enemy besides game developers.

Sure is a long way to 2011...

[panfist]

Here you go:
http://ixbtlabs.com/articles3/cpu/am...r2-533-p2.html

The biggest performance gain in applications: Unreal Tournament 3 (22%), World in Conflict (10%), 7-Zip (10%).

In this bench, they show DDR2 533 and 1066. DDR2 1066 is twice as fast as 533, and you get on average 3% gain.

So for a 100% increase in RAM speed...you get a 22% increase in one game. That also happens to be in a game that significantly favors the Intel architecture in the first place... World In Conflict hardly counts in your favor. A 10% return for a 100% increase in speed is not that impressive. Again, i7 will blow Phenom out of the water in this game anyway because of it's multithreading.

7zip...doesn't really count in AMD's favor either because it's far behind Intel. So what if double the RAM speed gets you a 10% gain when spending the same amount of money on an Intel system will net you a 25% gain?

HardOCP's position is that the jump from 800 to 1066 MHz RAM would not provide a significant performance increase. You are supporting HardOCP's position, just in case you didn't realize.

edit: looks like I'm not the first to come to the same conclusion.

[Shintai]

Your link has little to do with the subject, and my own link shows a good improvement in some benchmarks as memory speed increases.

The link is relevant to memory performance. Your link also shows the opposite. Guess you didnt like those results...

Average performance gain - 3%

You had to pick the abnormalities just to even try to make a case.

DDR2-533 to DDR2-1066 is 100%. DDR2-800 to DDR2-1066 is 33%. Plus the lowered scaling issue. Even by the best we talk 1% average. or less than 2% in Crysis. Thats with 100% scaling.

[roofsniper]

thanks for the link shintai. really a lot of people are being fanboys about the whole ddr2 ddr3 thing. ddr3 really doesn't have that big of an advantage over ddr2. it gives some extra performance but it doesn't equate to much. sure ddr3 will be nice whenever it drops in price but right now i find it as pointless. i hope next month when there are am3 phenoms out you guys will just drop this because im not seeing that huge of a gain from ddr3. phenom II am2+ supports ddr2 1066 while the am3 ones will support ddr3 1333. ill give it 1% performance bonus max. remember the timings are going to go up as well when with ddr2 800 and 1066 the timings are typically the same.

[demonkevy666]

thanks for the link shintai. really a lot of people are being fanboys about the whole ddr2 ddr3 thing. ddr3 really doesn't have that big of an advantage over ddr2. it gives some extra performance but it doesn't equate to much. sure ddr3 will be nice whenever it drops in price but right now i find it as pointless. i hope next month when there are am3 phenoms out you guys will just drop this because im not seeing that huge of a gain from ddr3. phenom II am2+ supports ddr2 1066 while the am3 ones will support ddr3 1333. ill give it 1% performance bonus max. remember the timings are going to go up as well when with ddr2 800 and 1066 the timings are typically the same.

all the AM3 boards I see they'll at HT3.0 and 3.1 isn't comming out till what march ?
I hope Am3 chips have higher NB support. I've gotten mine up to 2,730mhz any higher and it won't post.
I can set my HT to 2.6ghz and Nb to 2.6ghz and at 1.350 volts for nb.

[Shintai]

thanks for the link shintai. really a lot of people are being fanboys about the whole ddr2 ddr3 thing. ddr3 really doesn't have that big of an advantage over ddr2. it gives some extra performance but it doesn't equate to much. sure ddr3 will be nice whenever it drops in price but right now i find it as pointless. i hope next month when there are am3 phenoms out you guys will just drop this because im not seeing that huge of a gain from ddr3. phenom II am2+ supports ddr2 1066 while the am3 ones will support ddr3 1333. ill give it 1% performance bonus max. remember the timings are going to go up as well when with ddr2 800 and 1066 the timings are typically the same.

One good thing about AM3 tho is that the NB gets a 200Mhz increase. So it could give some boost there. Unfortunately it might also be a false marketing for the memory.

all the AM3 boards I see they'll at HT3.0 and 3.1 isn't comming out till what march ?
I hope Am3 chips have higher NB support. I've gotten mine up to 2,730mhz any higher and it won't post.
I can set my HT to 2.6ghz and Nb to 2.6ghz and a.350 volts for nb.

HT wont do anything absicly. For NB speed look above.

[Miss Banana]

The link is relevant to memory performance.

Of intel platforms yes, and were we talking about ddr3 on intel platforms?
No Shintai, we were talking about ddr2 on AMD platforms.

Your link also shows the opposite. Guess you didnt like those results...

Average performance gain - 3%

You had to pick the abnormalities just to even try to make a case.

DDR2-533 to DDR2-1066 is 100%. DDR2-800 to DDR2-1066 is 33%. Plus the lowered scaling issue. Even by the best we talk 1% average. or less than 2% in Crysis. Thats with 100% scaling.

Sorry how is a 3% gain the opposite of what I was claiming?
I was claiming there was an improvement, and that the improvement was especially substantial in some games and winzip.

I really wonder when you will be making your point...
Maybe you allready made your point?
In that case it wasn't a very good one.

[demonkevy666]

I have yet to see a comparsion of NB speed increase % from deneb.
agena's was 0.40-2.8%increase form 200mhz.


http://www.xbitlabs.com/articles/cpu...x4-9850_4.html

[Shintai]

Of intel platforms yes, and were we talking about ddr3 on intel platforms?
No Shintai, we were talking about ddr2 on AMD platforms.



Sorry how is a 3% gain the opposite of what I was claiming?
I was claiming there was an improvement, and that the improvement was especially substantial in some games and winzip.

I really wonder when you will be making your point...
Maybe you allready made your point?
In that case it wasn't a very good one.

Thats so funny. Nothing better than denial. Take and look on Crysis. Its the only thing that goes across. Add your 2% or actually even less. Tell us how that works out for the scores.

You picked 2 abnormalities and claims its like a standard.

[roofsniper]

all the AM3 boards I see they'll at HT3.0 and 3.1 isn't comming out till what march ?
I hope Am3 chips have higher NB support. I've gotten mine up to 2,730mhz any higher and it won't post.
I can set my HT to 2.6ghz and Nb to 2.6ghz and at 1.350 volts for nb.

idk if HT 3.1 is ever coming to phenom II. HT 3.0 is everything up until 2.6ghz and considering we are only at 1.8 right now on phenom II and 2.0 is coming soon i doubt we will see 2.6ghz+. just because it has been invented doesn't mean it will be implemented. if i remember correctly HT 3.0 was created in 2006 but cpus didn't use that right away.

[Donnie27]

64bit doesnt use it since they already use 40-48bit addressing or so. PAE is simply 36bit hardware addressing.

PAE is on by default, hence not in the boot.ini file.

There is also no slowdown with PAE. Unless you use AWE applications like SQl. But the benefit is bigger than the penalty.

Thanks!

[marten_larsson]

Even if the performance gain is less than 0,1% it does not justify the difference in setups! No one would EVER conduct a serious research with such lacking method, or at least not any serious research. If the memories don't have to be the same speed or timings, why bother with the same GFX? Or PSU? You use as standardised equipment as possible to reduce number of potential errors. It's just stupid to defend his choice of hardware - no one would ever "allow" a PII to outrun an i7 simply because the first uses a 4870X2 while the latter has a 8600GS and 512MB of RAM in a review and not call it biased (because it would have been, just as this one).

If he didn't get the 1066 memory to work on the PII, why not using the same 800 on the C2Q? If the performance loss/gain is close to nothing - what sense does it make to make them different?! If one wants to be taken seriously, the reviews have to have at least a bit scientific approach. This one clearly lacks quite many of those requirements.